Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus...

8
Embryotoxicity of TiO 2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, * , Diego Minetto a , Sara Totaro b , Ivan Mi ceti c b , Andrea Pigozzo b , Enrico Sabbioni b , Antonio Marcomini a , Annamaria Volpi Ghirardini a a Department of Environmental Sciences, Informatics and Statistics, University Cà Foscari Venice, Campo della Celestia, 2737/b, 30122 Castello, Venice, Italy b ECSIN e European Center for Sustainable Impact of Nanotechnology, Veneto Nanotech S.C.p.A., 45100 Rovigo, Italy article info Article history: Received 28 May 2013 Received in revised form 23 August 2013 Accepted 29 August 2013 Keywords: Mussel nTiO 2 Embryotoxicity Nanoecotoxicology abstract Few data exist on the ecotoxicological effects of nanosized titanium dioxide (nTiO 2 ) towards marine species with specic reference to bivalve molluscs and their relative life stages. Mytilus galloprovincialis Lamarck was selected to assess the potential adverse effects of nTiO 2 (0e64 mg/L) on its early larval development stages (pre-D shell stage, malformed D-shell stage and normal D-shell stage larvae) considering two exposure scenarios characterised by total darkness (ASTM protocol) and natural photoperiod (light/dark). This approach was considered to check the presence of potential effects associated to the photocatalytic properties of nTiO 2 . Parallel experiments were carried on with the bulk reference TiCl 4 . The toxicity of nTiO 2 showed to be mainly related to its nanocondition and to be inuenced by the exposure to light that supported the increase in the number of pre-D shell stage (retarded) larvae compared to the malformed ones especially at the maximum effect concentrations (4 and 8 mg nTiO 2 /L). The non-linear regression toxicity data analysis showed the presence of two EC50 values per exposure scenario: a) EC(50) 1 ¼ 1.23 mg/L (0.00e4.15 mg/L) and EC(50) 2 ¼ 38.56 mg/L (35.64 e41.47 mg/L) for the dark exposure conditions; b) EC(50) 1 ¼ 1.65 mg/L (0.00e4.74 mg/L) and EC(50) 2 ¼ 16.39 mg/L (13.31e19.48 mg/L) for the light/dark exposure conditions. The potential impli- cation of agglomeration and sedimentation phenomena on ecotoxicological data was discussed. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction Engineered nanomaterials are at the forefront of ecotoxicologist agendas due to their increasing use in a broad range of industrial and consumer products. Actually, they are manufactured in increasing amounts year-by-year (Ward and Kach, 2009). Particu- larly, nanosized titanium dioxide (nTiO 2 ) is used in a variety of in- dustries mainly for catalysis and photocatalysis applications and as an additive in paints, papers, inks, plastics and various consumer products (Menard et al., 2011). These extensive applications have led to a rapid increase in its production rate, which is estimated to reach in the United States 2.5$ 10 6 tons per year by 2025 (Jemec et al., 2008; Robichaud et al., 2009). It is reasonable to assume that such widespread use of nTiO 2 will result in an increased environmental exposure to titania (Hall et al., 2009) that may reach concentrations in surface waters posing a potential threat to aquatic ecosystems (Kaegi et al., 2008; Gottschalk et al., 2009; Scown et al., 2010; Wise and Brasuel, 2011). Most literature on the ecotoxicity of nTiO 2 deals with aquatic organisms such as bacteria, algae, invertebrates, and sh, but there are also case studies considering cell lines and rodents (Zhu et al., 2011a). Mostly, testing species are from freshwaters (Cattaneo et al., 2009). Little is known concerning salt water species such as mollusc bivalves. Indeed, amongst all environmental quality status sentinels, marine bivalves are considered a promising group of bio- indicators (OECD, 2010) and may represent a unique target group for nanoparticle toxicity with a special focus on Mytilus spp. (Canesi et al., 2012; Kadaret al., 2011, 2012). Essentially, they have most of the characteristics that a good bioindicator should have. They are substantially widespread in all fresh, brackish and salt water en- vironments, having a key ecological role both as lter and deposit feeders according to various species and habitats (His et al., 1997; Kadar et al., 2012). All stages of their life cycles (gametes, em- bryos, larvae and adults) have been used to dene an endpoint for a specic monitoring purpose (acute, sub-chronic or chronic toxicity). Bivalve molluscs toxicity tests are generally easy to perform, highly sensitive, cost-effective and not time consuming, ranging from 24 h to 48 h to obtain a rst meaningful result, or even *Corresponding author. Department of Environmental Sciences, Informatics and Statistics, University Cà Foscari Venice, Calle Larga Santa Marta, 2137, 30123 Dor- soduro, Venice, Italy. Tel.: þ39 041234 7737/8596, þ39 041234 7747/8929. E-mail address: [email protected] (G. Libralato). Contents lists available at ScienceDirect Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev 0141-1136/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.marenvres.2013.08.015 Marine Environmental Research 92 (2013) 71e78

Transcript of Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus...

Page 1: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

lable at ScienceDirect

Marine Environmental Research 92 (2013) 71e78

Contents lists avai

Marine Environmental Research

journal homepage: www.elsevier .com/locate/marenvrev

Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk)

Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro b, Ivan Mi�ceti�c b, Andrea Pigozzo b,Enrico Sabbioni b, Antonio Marcomini a, Annamaria Volpi Ghirardini a

a Department of Environmental Sciences, Informatics and Statistics, University Cà Foscari Venice, Campo della Celestia, 2737/b, 30122 Castello, Venice, Italyb ECSIN e European Center for Sustainable Impact of Nanotechnology, Veneto Nanotech S.C.p.A., 45100 Rovigo, Italy

a r t i c l e i n f o

Article history:Received 28 May 2013Received in revised form23 August 2013Accepted 29 August 2013

Keywords:MusselnTiO2

EmbryotoxicityNanoecotoxicology

*Corresponding author. Department of EnvironmenStatistics, University Cà Foscari Venice, Calle Larga Sasoduro, Venice, Italy. Tel.: þ39 041234 7737/8596, þ3

E-mail address: [email protected] (G. Lib

0141-1136/$ e see front matter � 2013 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.marenvres.2013.08.015

a b s t r a c t

Few data exist on the ecotoxicological effects of nanosized titanium dioxide (nTiO2) towards marinespecies with specific reference to bivalve molluscs and their relative life stages. Mytilus galloprovincialisLamarck was selected to assess the potential adverse effects of nTiO2 (0e64 mg/L) on its early larvaldevelopment stages (pre-D shell stage, malformed D-shell stage and normal D-shell stage larvae)considering two exposure scenarios characterised by total darkness (ASTM protocol) and naturalphotoperiod (light/dark). This approach was considered to check the presence of potential effectsassociated to the photocatalytic properties of nTiO2. Parallel experiments were carried on with the bulkreference TiCl4. The toxicity of nTiO2 showed to be mainly related to its “nano” condition and to beinfluenced by the exposure to light that supported the increase in the number of pre-D shell stage(retarded) larvae compared to the malformed ones especially at the maximum effect concentrations (4and 8 mg nTiO2/L). The non-linear regression toxicity data analysis showed the presence of two EC50values per exposure scenario: a) EC(50)1 ¼ 1.23 mg/L (0.00e4.15 mg/L) and EC(50)2 ¼ 38.56 mg/L (35.64e41.47 mg/L) for the dark exposure conditions; b) EC(50)1 ¼ 1.65 mg/L (0.00e4.74 mg/L) andEC(50)2 ¼ 16.39 mg/L (13.31e19.48 mg/L) for the light/dark exposure conditions. The potential impli-cation of agglomeration and sedimentation phenomena on ecotoxicological data was discussed.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Engineered nanomaterials are at the forefront of ecotoxicologistagendas due to their increasing use in a broad range of industrialand consumer products. Actually, they are manufactured inincreasing amounts year-by-year (Ward and Kach, 2009). Particu-larly, nanosized titanium dioxide (nTiO2) is used in a variety of in-dustries mainly for catalysis and photocatalysis applications and asan additive in paints, papers, inks, plastics and various consumerproducts (Menard et al., 2011). These extensive applications haveled to a rapid increase in its production rate, which is estimated toreach in the United States 2.5$106 tons per year by 2025 (Jemecet al., 2008; Robichaud et al., 2009). It is reasonable to assumethat such widespread use of nTiO2 will result in an increasedenvironmental exposure to titania (Hall et al., 2009) that may reachconcentrations in surface waters posing a potential threat to

tal Sciences, Informatics andnta Marta, 2137, 30123 Dor-9 041234 7747/8929.ralato).

ll rights reserved.

aquatic ecosystems (Kaegi et al., 2008; Gottschalk et al., 2009;Scown et al., 2010; Wise and Brasuel, 2011).

Most literature on the ecotoxicity of nTiO2 deals with aquaticorganisms such as bacteria, algae, invertebrates, and fish, but thereare also case studies considering cell lines and rodents (Zhu et al.,2011a). Mostly, testing species are from freshwaters (Cattaneoet al., 2009). Little is known concerning salt water species such asmollusc bivalves. Indeed, amongst all environmental quality statussentinels, marine bivalves are considered a promising group of bio-indicators (OECD, 2010) and may represent a unique target groupfor nanoparticle toxicity with a special focus onMytilus spp. (Canesiet al., 2012; Kadar et al., 2011, 2012). Essentially, they have most ofthe characteristics that a good bioindicator should have. They aresubstantially widespread in all fresh, brackish and salt water en-vironments, having a key ecological role both as filter and depositfeeders according to various species and habitats (His et al., 1997;Kadar et al., 2012). All stages of their life cycles (gametes, em-bryos, larvae and adults) have been used to define an endpoint for aspecific monitoring purpose (acute, sub-chronic or chronictoxicity). Bivalve molluscs toxicity tests are generally easy toperform, highly sensitive, cost-effective and not time consuming,ranging from 24 h to 48 h to obtain a first meaningful result, or even

Page 2: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e7872

lower when some biomarker endpoints are taken into consider-ation. The main disadvantage is related to the reproductive period,that could be quite short for some species sampled from the wildfor the performance of embryotoxicity tests. However, conditioningtechniques exist and may help to provide reproductive organismsthroughout all the year (Libralato et al., 2010; Mamindy-Pajanyet al., 2010; OECD, 2010; Canesi et al., 2012).

In vivo experiments with Mytilus galloprovincialis Lamarck werecarried on considering various biomarker endpoints after 24-hexposure to nTiO2. Results showed that nTiO2 (22 nm averageparticle size, 51m2/g) induced lysosomalmembrane destabilisationin the haemocytes and digestive gland (Canesi et al., 2010). It wasalso found that nTiO2 stimulated an increase in lysosomal lipofuscinand the lysosomal accumulation of neutral lipids, as well as theenhancement of the activity of catalase and glutathione transferasein the digestive glands. No effect on catalase and glutathionetransferase activities in the gills and no mortality were observed(Canesi et al., 2010).

Recently, it has been reported that nTiO2 produced no significantchanges in viability ofM. galloprovincialis hemocytes under variousexperimental conditions (1, 5 and 10 mg/L) (Ciacci et al., 2012),whereas lysosomal destabilisation occurred at 5 and 10 mg/mLwith values of �18 and �39%, respectively. Moreover, it was stim-ulated the lysozyme release at both concentrations, increasingphagocytosis of Neutral Red-conjugated zymosan particles at 5 mg/mL and strongly decreasing the phagocytic activity at 10 mg/mL.The N-oxides related effects evaluated as cyt c reduction werecomparatively higher with nZnO than with nTiO2. The assessmentof nitric oxide (NO) production indicated that 1 mg nTiO2/ml wasineffective, whereas 5 mg nTiO2/ml produced a time-dependentincrease in NO production. The transmission electronic micro-scopy analysis (TEM) showed that nTiO2 did not apparently affectthe hemocytes morphology, even though, after 60 min, nano-particles were observed within the nucleus (Ciacci et al., 2012).Other authors (Zhu et al., 2011a, 2011b) assessed the acute and sub-chronic toxicity and oxidative stress on nTiO2 in marine abaloneHaliotis diversicolor supertexta Reeve. In particular, no develop-mental effects of nTiO2 were observed at 2 mg/L, but hatching in-hibition and malformations were observed at 10, 50 and 250 mg/L(Zhu et al., 2011a). It was observed that 2 mg/L nTiO2 seem topotentiate the toxicity effect of tributyltin (TBT) compared to theaction that it may exert alone (Zhu et al., 2011a). The hatching in-hibition effective concentration inducing a 50% effect (EC50) after10 h exposure was set at 56.9 mg/L (28.0e126.3 mg/L) and themalformation EC50 (8 h) at 345.8 mg/L (155.0e1592.6 mg/L),whereas the no observed effect concentration (NOEC) value forboth endpoints was about 2 mg/L (Zhu et al., 2011a). Moreover, itwas highlighted that after 96 h exposure of adults in semi-staticconditions (daily water change with redosing) no acute toxicitywas detected within 0.1e10 mg/L (Zhu et al., 2011b). However, theactivity of superoxide dismutase significantly increased in thegroup exposed to 1.0 mg/L nTiO2, whereas glutathione decreased intreatments presenting �1.0 mg/L of nTiO2. Furthermore, reactiveoxidative species (ROS) generated by illumination of nTiO2 (Jianget al., 2008) may contribute to disturb the anti-oxidant system(Brown et al., 2004), potentially damaging lipids, carbohydrates,proteins and DNA (Kelly et al., 1998) and disruption of intra-cellularmetabolic activities (Long et al., 2006).

The aim of this work was to assess the potential toxicity of nTiO2compared to its bulk reference (titanium tetrachloride, TiCl4)through the embryotoxicity on M. galloprovincialis (48 h-oldlarvae). A two-way experimental design was used considering ascenario including not only the natural photoperiod, but also totaldarkness (ASTM, 2004). Thus, data interpretation could take intoaccount the effects associated to both the nano-dimension as well

as the photocatalytic properties of nTiO2 (i.e. ROS generation inpresence of light) mimicking a more environmentally realisticresponse.

2. Materials and methods

Several sets of experiments were performed to explore theagglomeration and sedimentation behaviour of the test nano-material in artificial seawater (ASTM, 2004) within a 50 h moni-toring period at 0.01, 0.1, 1 and 10 mg/L of nTiO2. Data are shown inthe dedicated paper of Brunelli et al. (2013).

2.1. Preparation of nTiO2 stock suspensions

Nano-Titanium Dioxide (P25, declared purity >99%) fromDegussa Evonik (Darmstadt, Germany) was selected for theexperimental activities. Bulk titanium (TiCl4) from Fluka-Sigma-Adrich (CAS: 7550-45-0) was considered as the non-nano controlwithin our experiments.

Stock suspensions for toxicity testing were prepared in artificialseawater (ASTM, 2004) from a primary suspension of 1 g nTiO2/Land 1.73 gTiCl4/L. Due to the fact that the bulk titanium standardwas dispersed in concentrated HCl, treatment solutions werebuffered with NaOH 3 M to a pH value of about 8.30. All suspen-sions/solutions were sonicated singly (UPe100H, Hielscher Ultra-sound Technology, Teltow, Germany) for 20min at 100W cooling atthe same time the preparing dispersion in an ice bath. Stock sus-pensions/solutions were arranged at least 1 h before testing toallow them to equilibrate and stored in amber flasks in the dark andthermostated to the target temperature. Suspensions/solutionswere resuspended via 10 s vortex before transferring to the expo-sure vessels. The experimental design considered the followingnominal concentrations: 0.5, 1, 4, 8, 16, 32 and 64 mg/L for bothnTiO2 and TiCl4.

2.2. nTiO2 characterization

The nanoparticles characterisation (specific surface area, shape,average dimensions and hydrodynamic radius) considered a com-bination of analytical techniques such as Brunauer-Emmett-Teller(BET) particle sizer (Autosorb�-iQ, QuantaChrome Instruments),Transmission Electron Microscopy (Fei Tecnai T12), Dynamic LightScattering (DLS) (ZetaSizer Nano ZEN3600, Malvern).

For BET analysis, a sample of 0.7514 g of nTiO2 was outgassedconsidering a defined heating profile (target temperature (�C) e

rate of degree/minute e soak time in minutes) (50/5.00/15e100/5.00/15e150/5.00/15e200/5.00/15e250/5.00/15e300/5.00/600).

For TEM analysis (120 kV), nanoparticles were resuspended inultrapurewater at a nominal concentration of 1440mg/L, depositedon copper grids (10 mL) with a continuous carbon film coating andobserved only once the dispersant was completely evaporatedovernight at room temperature.

For DLS analysis, nTiO2 reconstituted seawater (ASTM, 2004)dispersions were assessed at 18.0 � 0.1 �C (i.e. toxicity testingtemperature) after 48 h ageing and 1 min hand shaking.

2.3. Toxicity test

Batches of M. galloprovincialis were collected along the Adriaticcoast (Venice, Italy) during the breeding season (April) and usedjust few hours after sampling. The embryotoxicity test was per-formed according to ASTM (2004) and Libralato et al. (2010).Adults were induced to spawn by alternated thermal stimulationcycles (18 � 1 �C and 28 � 1 �C). Artificial seawater prepared withanalytical grade reagents for gametes collection and embryos

Page 3: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

Fig. 1. TEM picture of nTiO2 (P25, Degussa Evonik, Germany) (150000x).

Table 2Results of ANOVA comparison between the ecotoxicological effects from dark andlight/dark exposure scenarios; concentrations are in mg/L.

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e78 73

testing was the ASTM one (ASTM, 2004) characterised by a salinityof 34 psu. During each test, gametes derived from a batch of threemales and three females were separately filtered at 32 mm (spermcells) and 100 mm (eggs) to remove impurities. Eggs suspended in500 mL artificial seawater were fertilized injecting an aliquot ofsperm suspension yealding a ratio of approximately 1: 1 �106 egg/sperm in the final mixture; fertilization success was qualitativelychecked by microscopy. Egg density was determined by countingfour subsamples of known volume. Fertilized eggs, added to thetest solutions in order to obtain a density of about 70 eggs/mLwithin a 3 mL final volume, were incubated for 48 h at 18 � 1 �C.At the end of the test, samples were fixed with buffered formalin(4%) and 100 larvae were counted per any replicate suspension,distinguishing between normal larvae (D-shell stage) and abnor-malities (malformed larvae e concave, malformed or damagedshell, protruding mantle e and pre-D stages e trochophore larvaeor earlier stages). Sample analysis was carried on via an invertedmicroscope (Leica, Germany) provided with an image acquisitiontoolbox at 400x enabling to take representative light microscopicpictures.

Sterile capped polystyrene 3 mL 24-well microplates (Iwakibrand, Asahi Techno Glass Corporation, Tokyo, Japan) were usedas test chambers for the toxicity test. The acceptability of testresults was based on negative controls (ASTM dilution water) fora percentage of normal D-shell stage larvae >70% (ASTM, 2004).A copper solution prepared from copper nitrate standard foratomic absorption spectroscopy in ASTM dilution water was usedas positive control considering as reference the followingacceptability range 9.47 mg/L � EC50 � 21.72 mg/L (Libralato et al.,2009).

Stock and testing suspensions of nTiO2 and solutions of TiCl4were prepared within the test day. All suspensions were sonicatedfor 20 min at 100 W with a UP200S Hielscher Ultrasonic Technol-ogy (Teltow, Germany) cooling the preparing dispersions in an icebath. Once required, light was produced by a set of fluorescentlamps (Philips longlife, 1200 lumen, six lamps in series) to simulatemore likely natural levels of sunlight (6000e10000 lux) consid-ering 16 h light/8 h darkness.

Table 1Trend of the size of nTiO2 (P25, Degussa Evonik, Germany) agglomerates in artificial sea

nTiO2 (mg/L) 0 0.5 1

Average size of agglomerates (nm)e 18.0 � 0.1 �C (n ¼ 3) e 217 � 24 25

2.4. Data analysis

Toxicity effect data were determined as percentages ofabnormal larvae (number of retarded and malformed larvae on 100counts). EC50 on the exposed populations have been provided aswell as the relative confidence limit values at 95%. The responsesfor each treatment were corrected for effects in the negative controlby applying Abbott’s formula (ASTM, 2004). The hypothesis testwas verified using Analysis of Variance (ANOVA) and Tukey’s test tocheck any difference among the groups after lognormal trans-formation of concentration data. When ANOVA revealed significantdifferences among treatments, post-hoc tests were carried on withDunnett’s method testing the pairwise difference between eachtreatment and the control. Parametric or non-parametric methodswere considered for points’ estimation. All results are presented asmeans � standard error using for all statistical analysis the default5% rejection level. Statistical analyses were performed usingXLSTAT (Addinsoft).

3. Results and discussion

3.1. NPs characterisation

The mean average size of primary particles (anatase/rutile 3:1)in artificial seawater (ASTM, 2004) after 48 h ageing in the darknesswas 24.00 � 9.55 nm (n ¼ 724, TEM) (Fig. 1) presenting a shapepartly irregular and semi-spherical. The average size of aggregatesmeasured in artificial seawater per every single treatment was inthe range 217e590 nm at 18.0 � 0.1 �C (DLS) as reported in detailsper every single treatment in Table 1. The nTiO2 specific surface areawas 4.478 m2/g (BET) (intercept ¼ 0.2919, r ¼ 0.9999,constant ¼ 2664.24); it was classified as microporous.

3.2. Embryotoxic effects of TiCl4

The negative controls for both dark (11% � 0% of effect) andlight/dark (12% � 2% of effect) scenarios complied with the stan-dard value (<30% of effect) (ASTM, 2004) and showed to be notsignificantly different (p < 0.05). Moreover, the reference toxicant

water (ASTM, 2004).

4 8 16 32 64

2 � 54 303 � 106 473 � 105 469 � 66 426 � 72 590 � 59

Page 4: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e7874

exhibited the expected effect generating an EC50 ¼ 10.31 mg/L(9.35e11.43 mg/L) (Probit analysis) falling within the range reportedin Libralato et al. (2009). According to Fig. 2, all tested concentra-tions in both experimental conditions presented a percentage ofnormally developed D-shell stage larvae averagely equal or greaterthan 79%. Toxicity effects are mainly related to the presence ofretarded larvae (6e17%), while the amount of malformed D-shellstage larvae is rather small (2e6%). This suggests that toxicity ef-fects occurred at an early stage of the embryo larval developmentafter the zygote formation. No EC50 values were determined andthe maximum effects were evidenced at 64 mg/L in both scenariosas showed in Fig. 2. In particular, experiments carried on in totaldarkness showed a maximum effect of 20% � 4% (n ¼ 3), whereasthe light/dark scenario generated a maximum effect of 21% � 5%(n ¼ 3). They are significantly different from the negative controls(p > 0.05), but the difference between them is not consideredsignificant (p < 0.05). Thus, the presence of light during the 48 htest with mussel embryos exposed to TiCl4 treatments does notsignificantly affect toxicity compared to its total absence. Moreover,considering the adjustment for the effects within the negativecontrols according to Abbott’s formula, the net maximum adverseeffects are 12% � 4% (dark) and 14% � 5% (light/dark), respectively.

Fig. 2. Trend of TiCl4 toxicity data in both dark and light/dark exposure scenarios showingstage and pre-D shell stage (retarded) larvae to the final toxicity.

3.3. Embryotoxic effects of TiO2 nanoparticles

The effects in the negative controls for both dark (12% � 4%,n¼ 3) and light/dark (16%� 2%, n ¼ 3) scenarios complied with thestandard value (<30% of effect) (ASTM, 2004) and were notsignificantly different (p < 0.05). Moreover, the reference toxicantshowed the expected effect generating an EC50 ¼ 11.35 mg/L(10.57e12.19 mg/L) (Probit analysis) falling within the range re-ported in Libralato et al. (2009).

Toxicity data for both scenarios were presented in Fig. 3including details about the contributions to the final toxicity dueto the amount of malformed D-shell stage larvae and retarded pre-D shell stage larvae compared to normally developed ones. Thetrend of adverse effects is similar in both experimental conditionsevidencing that the maximum effects occurred at 4 and 8 mg/L.These concentrations presented a prevailing amount of malformedD-shell stage larvae rather than retarded larvae (pre-D shell stage).Nevertheless toxicity effects might have occurred since the begin-ning of the exposure, they did not substantially impaired the em-bryo larval development, but induced the generation of malformedlarvae after the first metamorphosis from the trochophore stagelarva to the D-shell one. Particularly, the amount of retarded larvae

in details the contributions as percentage of normal D-shell stage, malformed D-shell

Page 5: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

Fig. 3. Trend of nTiO2 toxicity data in both dark and light/dark exposure scenarios showing in details the contributions as percentage of normal D-shell stage, malformed D-shellstage and pre-D shell stage (retarded) larvae to the final toxicity.

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e78 75

at 4 and 8 mg/L is more than double in the light/dark scenariocompared to the dark one (p < 0.05) suggesting that the exposureto light increased the adverse effects of nTiO2.

The reason why the maximum adverse effects were detected inboth scenarios at 4 and 8 mg/L remained unclear. Actually, it couldbe suspected that the specific dimensional range of nTiO2 aggre-gates at 4 and 8 mg/L is in some way more bioavailable than thoseof the other treatments and appeared to be maximised at303 � 106 nm and 473 � 105 nm (Table 1), respectively. Anyway,this is a high speculative explanation that will requiremore focusedexperiments due to both the limits of the DLS technique (tendencyto data overestimation, need of highly stable suspesions and arelatively high number of particles per unit volume for a repre-sentative output) and the fact that nTiO2 agglomerates similar insize (16 and 32 mg/L, Table 1) to the above mentioned ones led toonly minor toxicity effects in both scenarios.

In the dark exposure scenario, the ANOVA evidenced that theeffects on negative controls always differ from treatments exceptfor 0.5 and 1 mg/L (p < 0.004). Considering the differences withintreatments, they always differ from each other except for 0.5e1, 4e8 and 16e32 mg/L (p > 0.05). Moreover, in the light/dark exposurescenario the ANOVA showed that the effects on negative controlsalways differ from treatments except for 0.5 and 64 mg/L

(p < 0.001). Considering the differences within treatments, theyalways differ from each other except for 0.5e32, 0.5e64, 4e8 and32e64 mg/L (p > 0.05). Considering the comparison between darkand light/dark exposure scenarios, the two groups of negativecontrols do not significantly differ from each other (p < 0.01).Similarly, it occurred for the assessment within treatments assummarised in Table 2 for 0.5e0.5, 0.5e32, 0.5e64, 1e0.5, 1e32, 1e64, 4e4, 4e8, 8e4, 8e8 and 64e1 mg/L (p < 0.05) for dark andlight/dark exposure scenarios, in that order.

Besides, the maximum effects observed for both scenarios at 4and 8 mg/L, the lowest effects were detected at the 0.5 mg/L ofnTiO2 for both dark and light/dark experiments presenting anaverage effect of 14.33% � 4.04% and 9.67% � 2.08% in that order.Moreover, further low toxicity effects were detected at 32 mg/L ofnTiO2 within the light/dark experiments (3.33% � 5.77%) and at64 mg/L of nTiO2 within the dark scenario (28.67% � 10.50%).

Toxicity data have been plotted in Fig. 4 as toxicity frequencieswithin the 0e1 range against the exposure concentrations aftertheir lognormal transformation in order to highlight the resultsobtainted at the lowest tested concentrations (0.5 and 1 mg/L).Each data set was analysed for the best fit. Only non-linearregression analysis provided suitable correlation coefficients (R2)of about 0.80 as shown in Fig. 4. As a consequence, two likely EC50

Page 6: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

Fig. 4. Toxicity data distribution as percentage of effect (abnormal and retarded larvae)as frequency plotted against nTiO2 concentration after lognormal transformation forboth dark and light/dark exposure scenarios. The two types of dashed lines describethe potential modeled trend of the two datasets.

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e7876

solutions may be obtained from each equation for both exposurescenarios: a) EC(50)1 ¼ 1.23 mg/L (0.00e4.15 mg/L) andEC(50)2 ¼ 38.56 mg/L (35.64e41.47 mg/L) for the dark exposurescenario; b) EC(50)1 ¼ 1.65 mg/L (0.00e4.74 mg/L) andEC(50)2 ¼ 16.39 mg/L (13.31e19.48 mg/L) for the light/dark expo-sure scenario. This fact suggests that hazard assessment of nTiO2

should take care of considering primarily the lowest EC50 valueaccording to the precautionary principle. It can be observed thatthere is apparently no significant difference (p < 0.05) between the

Fig. 5. Normally developed and abnormal (malformed or retarded) M. galloprovincialis larvn ¼ normal larva; p ¼ protruded-mantle larva; r ¼ retarded larva e trochophore stage.

EC(50)1 values from both scenarios. Conversely, the EC(50)2 ofnTiO2 from the light/dark scenario is lower than the EC(50)2 of thedark scenario, meaning that exposure to light might have increasedthe toxicity effect of nTiO2. Nevertheless, two exceptions werefound, apart the EC50 values determined on the basis of the best fitapproach. Indeed, the effects detected within the light/dark sce-nario at 32 and 64 mg/L, the highest tested concentrations, weresignificantly lower (p < 0.05) than the dark one without any clearexplanation at the moment. According to the ecotoxicological re-sults obtained for the bulk reference TiCl4 in both exposure sce-narios, most part of the adverse effects of nTiO2 can be attributed toit “nano” condition. Indeed, the effects of TiCl4 adjusted for theeffects in the control via the Abbott’s formula are 4% � 2% and11%� 5% at 4 mg/L and 10%� 1% and 6%� 4% at 8 mg/L for the darkand light/dark scenarios, in that order.

The existence for M. galloprovincialis of non-linear responses inthe recorded toxicity effects may be due to a wide series of eventsrelated to the exposure to nTiO2 that cannot be adequately inves-tigated in the present paper starting from direct/indirect cellentrance at an early embryo larval stage to ingestion rate andegestion dynamics of nTiO2 starting from the D-shell larvae stage.Within all these scenarios, the presence of light showed to signif-icantly influence the ecotoxicological effects increasing the numberof retarded larvae compared to malformed ones at least in coinci-dence to the maximum effect concentrations (4 and 8 mg/L).

Moreover, ecotoxicological effects can be potentially related tothe level of bioavailability of nanomaterials and thus to theiragglomeration and sedimentation behaviour in synthetic seawater(ASTM, 2004). Aggregation may influence the pathway of trophictransfer of nanoparticles as already suggested by Alber and Valiela

ae after 48 h from fertilization exposed to the 0.5, 1, 4, 8, 16, 32 and 64 mg/L nTiO2;

Page 7: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e78 77

(1994) and Kach and Ward (2008) for bivalves in their adult stage.Apparently, no data are available for bivalves ingestion rate andegestion dynamics during their larval stage such as the D-shellstage considered in the present research. A similar trend in toxicityresponse was observed by Spangenberg and Cherr (1996) onMytilus californianus (Conrad) larvae exposed to BaSO4. Indeed,BaSO4 above certain concentrations occurs as an insoluble precip-itate and thus is no more bioavailable in seawater highlighting thatlower level of Ba in seawater appear to result in greater toxicitythan do higher levels. Similarly, it might occur in relation to the sizeof nTiO2 aggregates where the rate of aggregation might control itsbioavailability. Indeed, Brunelli et al. (2013) explored the agglom-eration and sedimentation behaviour of nTiO2 (P25, Degussa Evo-nik, Darmstadt, Germany) in dispersions prepared according to thesame protocol used in this work over an observation time of 50 h.The results showed that a fast agglomeration takes place in theconcentration range 0.01, 0.1, 1 and 10 mg/L, followed by sedi-mentation of agglomerates. Extent and rate of nTiO2 agglomerationand sedimentation depended mostly on its initial concentrationand only to a minor extent on salt content, ionic strength, anddissolved organic carbon (Brunelli et al., 2013). Substantially, theactual exposure concentration was lower than the nominal con-centration and decreased over time (50 h) by a factor from 2 to 20,approximatively. Although these results did not account for theconsequences on nTiO2 behaviour due to the presence of biota, theysuggested a potential underestimation of the toxicity effects andthe related parameters such as the effective concentration values(i.e. the adverse effect is observed at a concentration that isdifferent, i.e. lower, than the nominal one) within the simplifiedmodel scenario. This could be considered as a first step towards thecomprehension of how nTiO2 acts in seawater media, beingnecessary to understand also how the biotic component mightinfluence the way it is exposed to nTiO2 investigating for exampleresuspension phenomena due to larvae swimming activities oraggregation events enhanced by extracellular compounds releasedinto the contaminated aqueous media (i.e. mucous). Of course, thiskind of observation is highly linked to the single testing protocoltaken into consideration and to the relative testing species sug-gesting a potentially species-specific acting behaviour.

Fig. 5 summarised the best pictures available from both exper-imental conditions presented in Figs. 3 and 4. From Fig. 5, it can behighlighted the presence of a visible amount of sedimented ag-glomerates (i.e. after 48 h of exposure in artificial seawater)increasing from the upper left corner (negative control) to thebottom right one (64 mg/L). It can be appreciated that normal D-shell larvae are also present at very high exposure concentrations(16, 32 and 64 mg nTiO2/L) as also diplayed by the ecotoxicologicalresults of Figs. 3 and 4. Moreover, besides the presence of retardedor malformed larvae, it can be noticed in Fig. 5 the presence of ring-like bidimensional structures surrounding some larvae mostly atthe pre-D trochophore stage. In particular, they were found mainlyat 4 and 8 mg/L of nTiO2 in both dark and light/dark scenarios ac-counting for about 1e3% of the total amount of abnormalities. Thepresence of these ring-like structures might be the result of anincrease in the general embryo stress, causing its developmentdelay as well as activating defensive systems such as the secretionof mucus preventing or limiting the undesirable presence ofnanoparticles aggregates. Thus, the toxicity effects of nTiO2 mightalso result in an energy reallocation from growth and developmentto the support of defensive mechanisms.

4. Conclusions

This paper assessed for the first time the nTiO2 toxicity effects onmussel early larval stages considering dark and light/dark exposure

scenarios within the 0e64 mg/L concentration range in artificialseawater. In parallel, the same experiments were carried on withTiCl4 as the bulk reference. The toxicity of nTiO2 showed to bemainly related to its “nano” condition and to be influenced by theexposure to light that amplified the ecotoxicological effectsincreasing the number of retarded larvae compared to the mal-formed ones obtained in total darkness especially at 4 and 8mg/L. Anon-linear regression analysis fitted the obtained transformed datashowing that the adverse effects are similar or slightly differentfrom the negative controls at the lowest and highest tested con-centrations, whereas they were maximised at 4 and 8 mg nTiO2/L.Due to the second order of the curves, two pairs of EC50 wereidentified per every experimental condition. Even though, thereasonwhy the maximum ecotoxicological effects were detected inboth scenarios at 4 and 8 mg/L remained unclear, suggestions weremade about how the bioavailability of nTiO2 might be influenced bythe dimension of its agglomerates. This is an interesting topic thatshould be further assessed to provide not only the actualbioavailable exposure to nTiO2 in ecotoxicological tests, but also forunderstanding its seawater mobility and fate in the biota and theenvironment.

References

Alber, M., Valiela, I., 1994. Incorporation of organic aggregates by marine mussels.Mar. Biol. 121, 259e265.

ASTM, 2004. Standard Guide for Conducting Static Acute Toxicity Tests Starting withEmbryos of Four Species of Saltwater Bivalve Molluscs. E724e1998/2004.American Society for Testing and Materials.

Brown, D., Donaldson, K., Borm, P., Schins, R., Dehnhardt, M., Gilmour, P.,Jimenez, L.A., Stone, V., 2004. Calcium and ROS-mediated activation of tran-scription factors and TNF-a cytokine gene expression in macrophages exposedto ultrafine particles. Ame. J. Physiol. e Lung Cellular Mol. Physiol. 286, 344e353.

Brunelli, A., Pojana, G., Callegaro, S., Marcomini, A., 2013. Agglomeration andsedimentation of titanium dioxide nanoparticles (n-TiO2) in synthetic and realwaters. J. Nanopar. Res. 15, 1684.

Canesi, L., Ciacci, C., Vallotto, D., Gallo, G., Marcomini, A., Pojana, G., 2010. In vitroeffects of suspensions of selected nanoparticles (C60 fullerene, TiO2, SiO2) onMytilus hemocytes. Aqu. Toxicol. 96, 151e158.

Canesi, L., Ciacci, C., Fabbri, R., Marcomini, A., Pojana, G., Gallo, G., 2012. Bivalvemolluscs as a unique target group for nanoparticle toxicity. Mar. Environ. Res.76, 16e21.

Cattaneo, A.G., Gornati, R., Chiriva-Internati, M., Bernardini, G., 2009. Ecotoxicologyof nanomaterials: the role of invertebrate testing. Invertebrate Surv. J. 6, 78e97.

Ciacci, C., Canonico, B., Bilanicôv�a, D., Fabbri, R., Cortese, K., Gallo, G., Marcomini, A.,Pojana, G., Canesi, L., 2012. Immunomodulation by different types of N-oxidesin the hemocytes of the marine bivalve Mytilus galloprovincialis. PLoS ONE 7, 1e10.

Gottschalk, F., Sonderer, T., Scholz, R.W., Nowack, B., 2009. Modeled environmentalconcentrations of engineered nanomaterials (TiO2, ZnO, CNT, Fullerenes) fordifferent regions. Environ. Sci. Technol. 43, 9216e9222.

Hall, S., Bradley, T., Moore, J.T., Kuykindall, T., Minella, L., 2009. Acute and chronictoxicity of nano-scale TiO2 particles to freshwater fish, cladocerans and greenalgae, and effects of organic and inorganic substrate on TiO2 toxicity. Nano-toxicology 3, 91e97.

His, E., Seaman, M.N.L., Beiras, R., 1997. A simplification of the bivalve embryo-genesis and larval development bioassay method for water quality assessment.Water Res. 31, 165e172.

Jemec, A., Drobne, D., Rem�skar, M., Sep�ci�c, K., Ti�sler, T., 2008. Effects of ingestednano-sized titanium dioxide on terrestrial isopods (Porcellio scaber). Environ.Toxicol. Chem. 27, 1904e1914.

Jiang, J., Oberdorster, G., Elder, A., Gelein, R., Mercer, P., Biswas, P., 2008. Doesnanoparticle activity depend upon size and crystal phase? Nanotoxicology 2,33e42.

Kach, D.J., Ward, J.E., 2008. The role of marine aggregates in the ingestion ofpicoplankton-size particles by suspension-feeding molluscs. Mar. Biol. 153,797e805.

Kadar, E., Tarran, G.A., Jha, A.N., Al-Subiai, S.N., 2011. Stabilization of engineeredzero-valent nanoiron with Na-acrylic copolymer enhances spermiotoxicity.Environ. Sci. Technol. 45, 3245e3251.

Kadar, E., Dyson, O., Handy, R.D., Al-Subiai, S.N., 2012. Are reproduction im-pairments of free spawining marine invertebrates exposed to xerp-valentnano-iron associated with dissolution of nanoparticles? Nanotoxicology,1e9.

Kaegi, R., Ulrich, A., Sinnet, B., Vonbank, R., Wichser, A., Zuleeg, S., Simmler, H.,Brunner, S., Vonmont, H., Burkhardt, M., Boller, M., 2008. Synthetic TiO2

Page 8: Marine Environmental Research · 2017. 12. 18. · Embryotoxicity of TiO2 nanoparticles to Mytilus galloprovincialis (Lmk) Giovanni Libralato a, b, *, Diego Minetto a, Sara Totaro

G. Libralato et al. / Marine Environmental Research 92 (2013) 71e7878

nanoparticle emission from exterior facades into the aquatic environment.Environ. Poll. 156, 233e239.

Kelly, K., Havrilla, C., Brady, T., Abramo, K., Levin, E., 1998. Oxidative stress in toxi-cology: established mammalian and emerging piscine model systems. Environ.Health Perspec. 106, 375e384.

Libralato, G., Losso, C., Avezzù, F., Volpi Ghirardini, A., 2009. Influence of the salinityadjustment methods, salts and brine, on the toxicity of wastewater samples tomussel embryos. Environ. Technol. 30, 85e91.

Libralato, G., Volpi Ghirardini, A., Avezzù, F., 2010. How toxic is toxic? A proposal forwastewater toxicity hazard assessment. Ecotoxicol. Environ. Safety 73, 1602e1611.

Long, T.C., Saleh, N., Tilton, R.D., Lowry, G.V., Veronesi, B., 2006. Titaniumdioxide (P25)produces reactive oxygen species in immortalized brain microglia (BV2): impli-cations for nanoparticle neurotoxicity. Environ. Sci. Technol. 40, 4346e4352.

Mamindy-Pajany, Y., Libralato, G., Roméo, M., Hurel, C., Losso, C., VolpiGhirardini, A., Marmier, N., 2010. Ecotoxicological evaluation of Mediterraneandredged sediment ports based on elutriates with oyster embryotoxicity testsafter composting process. Water Res. 44, 1986e1994.

Menard, A., Drobne, D., Jemec, A., 2011. Ecotoxicity of nanosized TiO2. Review ofin vivo data. Environ. Poll. 159, 677e684.

OECD, 2010. Detailed Review Paper (DRP) on Molluscs Life-cycle Toxicity Testing,Series on Testing and Assessment N. ENV/JM/MONO, p. 121.

Robichaud, C.O., Uyar, A.E., Darby, M.R., Zucker, L.G., Wiesner, M.R., 2009. Estimatesof upper bounds and trends in nano-TiO2 production as a basis for exposureassessment. Environ. Sci. Technol. 43, 4227e4233.

Scown, T.M., van Aerle, R., Tyler, C.R., 2010. Review: do engineered nanoparticlespose a significant threat to the aquatic environment? Critic. Rev. Toxicol. 40,653e670.

Ward, J.E., Kach, D.J., 2009. Marine aggregates facilitate ingestion of nanoparticlesby suspension-feeding bivalves. Mar. Environ. Res. 68, 137e142.

Wise, K., Brasuel, M., 2011. The current state of engineered nanomaterials in con-sumer goods and waste streams: the need to develop nanoproperty-quantifiable sensors for monitoring engineered nanomaterials. J. Nanotechn.Sci. Appl. 4, 73e86.

Zhu, X., Zhu, J., Cai, Z., 2011a. TiO2 Nanoparticles in the marine environment: impacton the toxicity of tributyltin to abalone (Haliotis diversicolor supertexta) em-bryos. Env. Sci. Technol. 45, 3753e3758.

Zhu, X., Zhou, J., Cai, Z., 2011b. The toxicity and oxidative stress of TiO2 nano-particles in marine abalone (Haliotis diversicolor supertexta). Mar. Poll. Bull. 63,334e338.